Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
THE MECHANISM OF ALDEHYDE AND KETONE PHOTOLYSIS1 PHILIP A. LEIGHTON Department of Chemistry, Stanford University, California Received May 25, 1958 I. ABSORPTION SPECTRA All aldehydes and ketones show an absorption band in the near ultra- violet, extending roughly from about 3500 A. to below 2500A., absorption in which probably produces transitions in non-bonding electrons of the carbonyl group (27, 45~. The band shows, in general, the common phe- nomenon of structure at longer wave lengths, changing to a continuum at shorter wave lengths (9, 12, 13, 14, 16, 17, 18, 19, 25, 33, 53~. Only in the case of formaldehyde has rotational structure been definitely observed and, except for formaldehyde, the transition from structure to continuum is very gradual. In the lighter aldehydes there appears to be a region of diQuse bands or predissociation preceding the continuum and, in addition, the bands themselves appear to be underlaid with a continuum which gradually becomes stronger as wave length is decreased. Fluorescence, where observed, is most intense for absorption in the long wave length part of the band, but the observed limits of fluorescence do not necessarily agree with the observed limits of structure in absorption (14, 18, 19, 53~. The structure becomes less marked with increase in magnitude of the hydrocarbon residue, while among the three classes, ketones, saturated aldehydes, and unsaturated aldehydes, it is interesting to note that ketones show the least structure but the most fluorescence, while the unsaturated aldehydes show the most structure and the least fluorescence. A second and much stronger region of absorption begins at about 2000 A. (2300 A. for unsaturated aldehydes). In those cases which have been investigated this is found to consist of series of either diffuse or discrete bands (49, 53), some of which fit a Rydberg formula (36~. The bands are underlaid with faint continuous absorption and are followed at still shorter wave lengths (<1500 A.) by a strong continuum (14~. Photochemical investigations have been confined almost entirely to the near ultraviolet ~ Contribution No. 5 to the Third Report of the Committee on Photochemistry, National Research Council. 51
52 PHILIP A. LEIGHTON region, absorption in which produces two reactions, decomposition and polymerization. II. THE PRIMARY PROCESS IN DECOMPOSITION Discrete bands together with fluorescence indicate the production of a relatively long-lived molecule on absorption at longer wave lengths, while the appearance of diffuse bands and continua indicate a dissociation process at shorter wave lengths, although, as the different types of absorption overlap, the resulting processes must also overlap. The observation that the gaseous products of photolysis in the case of the lighter aldehydes and ketones consist chiefly of carbon monoxide and hydrocarbons or hydrogen has led to the conclusion that the bond or bonds adjacent to the carbonyl group are dissociated as the result of absorption. Supporting this are energetic considerations (51) and the demonstration by Pearson and his collaborators of the production of free methyl radicals during the photolysis of acetone and, to a lesser extent, of acetaldehyde (43) as well as of ethyl, propyl, and butyl radicals during the photolyses of diethyl and higher ketones (42, 44~. Referring to photolysis through the breaking of bonds adjacent to the carbonyl group as type I, three primary processes appear energetically possible (28~: (A) A dissociation into hydrocarbon (or hydrogen in the case of formal- dehyde) and carbon monoxide molecules in one step: ROCOCO + he ~ RARE +-CO(~Z or 3~) where Rat, R2 = an alkyl radical or a hydrogen atom. (B) The dissociation of a single RC bond to produce an alkyl and an acyl radical: R,R2CO + by , Rt + R2CO (C) The dissociation of both R C bonds simultaneously to give the corresponding radicals and normal (~2) carbon monoxide: R:R2CO + by ~ Ri + R2 + COME The decision as to the relative importance of these different dissociation processes constitutes one of the present problems in the photochemistry of aldehydes and ketones. The fact that aldehydes, RCHO, give predomi- - nantly a single hydrocarbon of composition RH, while mixed ketones give a mixture of three hydrocarbons R:R,, R:R2, and R2R2, led Norrish and Kirkbride (34) to the conclusion That the primary process is represented by process A for aldehydes and process C for ketones. More recent obser- vations, however, necessitate the modification of both of these suggestions. The isolation of diacetyl by Barak and Style (4), the demonstration by
ALDEHYDE AND KETONE PHOTOLYSIS 53 Spence and Wild (50) that, for absorption in the continuum, the C2H6/CO ratio in the product gas is considerably greater than unity, and that the difference is quantitatively accounted for by diacetyl formation, and the detection by Glazebrook and Pearson (15) of acetyl radicals in concentra- tion comparable to that of methyl radicals, all resulting from the photolysis of acetone at room temperatures, is convincing evidence that process B. rather than process C, must be concerned in ketone photolysis, at least in the case of acetone. The production of hydrogen (7, 24, 26), as well as of alkyl radicals (41, 42, 43, 44), and the existence of a chain at higher tem- peratures (1, 21, 26) all indicate that process B or process C must be con- cerned to some extent in aldehyde photolysis. The low stationary concen- tration of atomic hydrogen compared to that of methyl radicals (11, 40) is an indication that this dissociation occurs by process B. with the splitting of the C" C rather than the HC bond, but the existing evidence is inconclusive in this regard. The quantum yield of decomposition of the aldehydes at temperatures below 100°C. has been found to increase with decreasing wave length in all cases studied except that of formaldehyde (24, 25, 26, 35~. Accompanying the increase in quantum yield is a marked increase in the yield of hydrogen (7, 24, 269. These facts have been explained by Rollefson (48) on the basis of a competition, following absorption, between four possible paths which determine the fate of the activated molecule: (~) deactivation by fluorescence or collision, (2) dissociation into hydrocarbon and carbon monoxide by process A, (~) dissociation into radicals by process B or process C, and (~) reaction with another molecule or molecules to form a! polymer. At longer wave lengths, or in the region of banded absorption, processes 1 and 4 are predominant; with decreasing wave length, in agree- ment with the disappearance of structure and fluorescence, processes 2 and 3 become the more important, with 3 increasing more rapidly than 2. Rollefson estimates in the case of acetaldehyde that, at 3130 A., 90 per cent of the molecules dissociating do so by process 2 and 10 per cent by process 3, while at 2537 A. 75 per cent dissociate by process 2 and 25 per cent by process 3. The probability of dissociation into finished molecules compared to that of dissociation into radicals is thus 9:1 at 3130 A., de- creasing to 3 :1 at 2537 A. Although the magnitude of these ratios depends upon arbitrary assumptions as to the rates of the secondary reactions (26), the predominance of the dissociation into finished molecules, particularly at longer wave lengths, is in accordance with the smaller number of methyl radicals produced in acetaldehyde as compared with acetone (43), the small number of hydrogen atoms produced by absorption in the predissociation region as compared with the continuum in formaldehyde (39), and with the observation of Norrish and Bamford (30, 32) that, except as there is a reaction with the solvent, for dipropyl ketone type I photolysis is almost
54 PHILIP A. LEIGHTON completely inhibited in solution (owing presumably to primary recombina- tion of the radicals produced), while for isovaleraldehyde the solvent has relatively little effect. On the other hand, the report of Bowen and de la Praudiere (10) that the photolysis of acetaldehyde to yield gaseous prod- ucts is virtually completely inhibited in the pure liquid would indicate a dissociation into radicals as predominant, while the statement of Akeroyd and Norrish (1) that, in the chain photolysis of acetaldehyde in the presence of acetone, it snakes little difference which of these substances absorbs the light, suggests that dissociation into radicals is as efficient a process for acetaldehyde as for acetone. Little evidence is available as to whether A and B are distinct processes (48), or whether process A results from a rapid secondary reaction following process B (7), vie., RCHO +hv- >R + CHO >RH + CO In the former case the increase in B with decreasing wave length would arise from the changing potential energy of the excited molecule and the resultant change in relative probability of transition into different unstable states; in the latter case, it would arise from a more rapid separation of the radicals with increasing energy absorbed, with resultant change in the secondary reactions. By extending the latter point of view the data thus far discussed are capable of explanation entirely on the basis of a primary dissociation into free radicals, followed by appropriate secondary reactions (8, 20~. Thus, while the production of some carbon monoxide and a hy- drocarbon in the initial act more readily accounts for certain features of these reactions, there appears to be no fact which definitely requires such an act as a separate primary process. When studying the photolysis of methyl butyl ketone, Norrish and Appleyard (31) found that an entirely different type of decomposition occurred, which can best be described as a species of cracking the hydro- carbon chain: CH3 CH3CH2CH2CH2 CH3 CO + he > CH3CH:CH2 + CO CH3 Similar reactions, as judged from the decomposition products, have since been observed for a number of aldehydes and ketones containing an alkyl chain of three or more carbon atoms (3, 26, 29~. In each case the bond between the carbon atoms or and ~ to the carbonyl group is broken, and a hydrogen atom or proton migrates to the cc-carbon atom, leaving an olefin hydrocarbon and forming acetaldehyde or a methyl ketone. Following a suggestion of Norrish, this mode of decomposition may be referred to as
ALDEHYDE AND KETONE PHOTOLYSIS 55 type II. Judging from the data in table 1, type II decreases in importance with decreasing wave length, but increases in importance with increasing length of the hydrocarbon chain. Evidence that type II dissociation results in finished molecules is fur- nished by the observation of Glazebrook and Pearson (15) that free radicals are apparently produced only by type I decomposition, and by the fact that, in all cases, type II decomposition occurs without modification in solution (29, 32~. As suggested by Norrish, it appears that this is a true primary process involving some type of resonance between the excited carbonyl group and the or, ,B C C bond. TABLE 1 Relative renumbers of molecules decomposing by type I and type 1I photolyses, as deter- mined by the composition of the products CO3§POUND n-Butyraldehyde . Isovaleraldehyde . . Dipropyl ketone..... Methyl butyl ketone. CONDITIONS 3130 A.; 30°C. 2654 A.; 30°C. 2537 A; 30°C Full radiation of Hg arc; approxi- mately room temperature Full radiation of Hg arc; approxi- mately room temperature 277~2480 in.; 127°C. TYPE ~ per cent 90 97 100 47 37 13 TAPE II per cent 10 3 o 53 63 87 III. SECONDARY REACTIONS IN DECOMPOSITION That a chain reaction follows the photodissociation of acetaldehyde was demonstrated by Leermakers (21), who found that the quantum yield increases from less than unity at room temperature to values of 100 or more at temperatures around 300°C. A similar increase has been observed for formaldehyde (1~; it is less marked for the butyraldehydes (26) and absent for valeraldehyde (21, 23~. For acetaldehyde, in the temperature region of long chains, the variation in yield with pressure and intensity is given by the equation, after Leer- makers, _d(CH3CHO) = k~.Iabs + k2I~bs2. (CH3C0O) (1) Taking into account the demonstration by Allen and Sickman (2) that methyl radicals initiate a chain decomposition of acetaldehyde, Leermakers
PHILIP A. LEIGHTON proposed a mechanism for the photolysis similar to that of Rice and HE feld (46) for the thermal decomposition (table 2~. Reaction a2 was divided into steps HCO ~ H + CO and H + CH3CHO ~ H2 + CHICO by Leermakers and by Akeroyd and Norrish (1~. The lack of a carte. monoxide deficiency in the products favors this division; the low atom hydrogen concentration during the reaction and the greater energel stability of formyl radicals as compared with acetyl radicals do not. R action a7 was included by Leermakers to explain certain differences b tween his work and that of Allen and Sickman, but Akeroyd and Norri TABLE 2 Reactions following dissociation intoiree radicals of acetaldebyde and acetone a. ACETALDERYDE b. ACETONE (al) CH3CHO + ha ~ CH3 + HCO (CH3)2CO + ha ~ CH3 ~ CH3CO (1 (as) HCO + CH3CHO ~ H2 + CO ~ CH3CO (as) CH3 + CH3CHO ~ CH4 + CH3CO CH3 ~ (CH3)2CO ~ CH4 ~ CH2COCH: (1 (a4) CH3CO + M ~ CH3 + CO + M (1 (as) 2CH3CO ~ (CH3CO)2 (1 (as) 2CH3 ~ C2H ~(1 (a7) HCO + CH3CHO ~ CH4 + CO + HCO (as) CH3 + HCO ~ CHAD + CO CH3 + CH3CO ~ C2H6 + CO (1 prefer to eliminate this step. In any case, a rate law in agreement wil equation 1 is obtained. Eliminating reaction a7, temperature coefficien of the photolyses of formaldehyde and acetaldehyde assign acti~ratic energies of 16 Cal. to reaction a2 and 10 Cal. to reaction a3 (1~. Recent analyses of Blacet and Volman (8) show that at room temper tures the gaseous products of acetaldehyde photolysis consist entirely hydrogen, methane, and carbon monoxide, with no ethane or ethylen The Hz/CO ratio decreases toward zero with increasing wave length al: increasing temperature above 30°C. and with decreasing temperatu: below 30°C.; it decreases but slightly with increasing pressure. The results are in accordance with the Leermakers mechanism provided the reaction ad serves as the chief chain-breaking step at low temperature and indicate that, contrary to Leermakers' conclusion, reaction a3 is sti edective even at room temperature. Contrasting sharply with the aldehydes, the quantum yield of acetor decomposition remains at unity or less even up to 400°C. (1, 22~. ~ room temperature the sole products in quantity, as demonstrated by Spent
ALDEHYDE AND :ELETONE PHOTOLYSIS and Wild (50), are diacetyl, ethane, and carbon monoxide. 57 With increas- ing temperature the yield of diacetyl decreases, disappearing entirely above 60°C., although diacetyl itself is stable at this temperature, while in the gaseous products the C2H6/CO ratio approaches unity and consider- able quantities of methane appear. For absorption in the continuum ~ < 2900 A.) at room temperature, the C2H6/CO ratio increases with ace- tone pressure and with light intensity, approaching a constant value at high intensities. At 60°C., also for absorption in the continuum, the yield of methane increases with increasing acetone pressure and with decreasing intensity. These data may be accounted for on the basis of Rice and Herzfeld's thermal decomposition scheme (46, 47), as later applied by Leermakers (22) and Patat (38) to the high-temperature photolysis of acetone (table 2~. The analytical data of Spence and Wild are explained by an increase in importance of reactions be and b4 with increasing temperature. At 20°C. reaction b3 is negligible, while the observed C2H6/CO ratio of 1.6 to 2.5 means that from 50 to 75 per cent of the acetyl radicals are reacting by reaction b5 to form diacetyl. At 60°C. absence of diacetyl indicates that the acetyl radicals are reacting entirely by reaction b4, while the amount of methane produced shows that from 15 to 40 per cent (depending ore light intensity and pressure) of the methyl radicals are reacting by reac- tion b3. The thermal stability at room temperature and the instability at slightly higher temperatures, which are thus indicated for acetyl radicals, are verified by the report of Glazebrook and Pearson (15) that these radicals may be detected dur~ng the photolysis at room temperature but not at 60°C. The absence of a chain in acetone photolysis signifies that the CH~COCTIa radicals produced by reaction b3 are incapable of further reaction with the acetone. Their ultimate fate is unknown. This mechanism is in har- mony with Glazebrook and Pearson's observation that methyl radicals yield no diacetyl with acetone, and with the observation of Akeroyd and Norrish that, in acetone-acetaldehyde mixtures, there is neither inhibition or sensitization of the photodecomposition of one substance by the other, and that it makes little di~erence which compound absorbs the light. The importance of reaction b5 at room temperatures in the acetone de- composition ~ustifies its inclusion as the chain-breaking step in the low- temperature aldehyde photolysis. With the increase in rate of reaction 4 at higher temperatures, reaction 5 disappears and reaction 6 presumably becomes the chain-breaking step. Reaction b8 was considered necessary by Spence and Wild to account for the approach to a constant C2H6/CO ratio at high intensities. The absence of ethane in the decomposition products of acetaldehyde (8) shows that reaction b8 is negligible in acetal- dehyde photolysis. In the discussion thus far, no means have been provided to account for .
58 PHILIP A. LEIGHTON the low quantum yields, particularly of acetone and of the unsaturated aldehydes. The decomposition yields of crotonaldehyde and of acrolein are below 0.04 at room temperature (5, 6~. Although no direct measure- ments appear to have been made of the quantum yield of acetone photo- lysis below 56°C., at this temperature, with an incident intensity of ~ 104 ergs cm.-2 sec.-i of 3130 A., it varies from 0.17 at atmospheric pressure to 0.04 at 50 to 80 mm. of acetone (12~. These values were calculated on the basis of pressure change with the assumption of two molecules of gas per molecule decomposed. That the yield is much lower at room temperature is indicated by the observation of Taylor and Jungers (52) that the rate of carbon monoxide evolution from acetone at ~ 80 mm. in the presence of ethylene at 150 mm. and exposed to the full radiation of a hot mercury arc is only one-tenth as great at 25°C. as at 80°C. Four means of accounting for these low yields have been proposed: (a) The primary yield is unity, but the overall yield is reduced by a recom- laination of radicals (6, 20, 24), e.g., Rat + R2CO ~ ROCOCO where R2 = a hydrogen atom or an alkyl group. (b) A reorganization or distribution of the absorbed energy occurs over many internal degrees of freedom, with the result that the life of the activated molecule is greatly increased and the chance for dissociation before ultimate collisional stabi- lization occurs is correspondingly decreased (29, 33, 48~. An observed continuous absorption would in this case be "experimentally continuous", owing to the overlapping of a large number of closely spaced transitions. (c) In the majority of cases quantum yields have been based on rate of carbon monoxide evolution. In these cases any condensation reactions (5, 12) would reduce the apparent yield. (d) Deactivation by collision or fluorescence may be a contributing factor following absorption at longer wave lengths (e.g., acetone at 3130 A.) (129. Since theory b accounts for the low overall yield on the basis of a low primary yield, the question of chain length in the secondary reactions im- mediately becomes concerned. If the acetaldehyde photolysis at high temperatures proceeds entirely by reactions al, a2, as, a4, and an, the rate of decomposition will be d(CH3CHO) _ o~r ~ k:~ ~,73b"-CH3CHo dt YE abs. T (2) where ~ is the primary quantum yield (reaction al). The chain length, which will then be given by k3(CH3CHO) 2kI/2~1/2Ilb2 (3)
ALDEHYDE AND KETONE PHOTOL~'SIS 59 is inversely proportional to the square root of the primary yield. It follows that the absence of a chain in the high-temperature acetone photolysis cannot be inferred on the basis of the low quantum yield alone. The yield shows a high temperature coefficient from room temperature at least to 160°C., and the observed values may be due to a very inefficient primary process. The fact, however, that methyl radicals from other sources fail to initiate a chain in acetone is confirming evidence that no chain is con- cerned in the photolysis. The observation of Norrish and Kirkbride (35) that formaldehyde vapor at 110°C. is decomposed by absorption of 3650 A., which lies in the region of definite rotational structure, with a quantum yield almost as great as that for absorption in the continuum (0.7 versus 1.0) led these authors to the postulation that activated molecules may be dissociated by collision. Absence of atomic hydrogen (37) and the low energy of the absorbed quantum would both indicate a dissociation into finished molecules, H2CO + he (> 3000 A.) ~ H2CO* H2CO* + M - > He + CO + M The increase in quantum yield with pressure for absorption in the banded region (3130 A.) for both acetone (12, 19) and the butyraldehydes (26) has likewise been interpreted as indicating a dissociation induced by collisions. That such a phenomenon is not general is shown by the decrease in yield with increasing pressure for acetaldehyde at 3130A. Independence between yield and pressure at shorter wave lengths indicates that these efl ects are related to the production of activated molecules. The decrease in yield with increasing pressure for acetaldehyde may be interpreted as indicating collisional deactivation or the removal of activated molecules by polymerization. The high decomposition quantum yield for acetalde- hyde at 3342 A., which falls in the weak continuum between two bands, as compared with the low yield at 3130 A., which falls directly on a band, is additional evidence in support of this view (25~. In the case of acetone, data as to the nature of the process at 3130 A. are conflicting. The C2H6/CO ratio of unity indicates a dissociation into finished molecules (50), while the known production of methyl radicals at this wave length indicates a dissociation into radicals. The change in quantum yield with intensity (12, 25) is difficult to reconcile with collisional dissociation. The data are best explained on the basis of a coexistence of two primary proc- esses due to the overlapping of two types of absorption,- discrete absorp- tion, with the production of activated molecules which are either deacti- vated or dissociated into finished molecules by collision, and continuous absorption, with direct dissociation into radicals.
60 PHILIP A. LEIGHTON IV. POLYMERIZATION Relatively little attention has been given to the photopolymerization of aldehydes and ketones, but the data available indicate that two distinct processes are concerned. With change in wave length from the region of structure and fluorescence to the region of continuous absorption, the quan- tum yield of apparent polymerization uniformly decreases for acetalde- hyde (25), first decreases (3130 to 3020 A.) then increases (3020 to 2537 A.) for propionaldehyde (24) and n-butyraldehyde (26), and uniformly in- creases for isobutyraldehyde. At 3130 A. the yield increases at least with the first power of the pressure, while at 2654 and 2537 A. it may either increase (propionaldehyde) or be independent (acetaldehyde) of the pres- sure. The two processes thus suggested are (a) polymerization involving activated molecules and (b) polymerization induced by the radicals result- ing from photodissociation by absorption in the continuum. Process a becomes successively less important and process b more important in the series from acetaldehyde to isobutyraldehyde. A remarkable increase in the rate of polymerization with decreasing wave length has been observed by Blacet, Fielding, and Roof (5) for acro- lein, the Quantum yield at P = 200 mm. and T = 30°C. changing from 0.3 at 3660 A. (structure) to ~ 20 at 2654-2537 A. (continuum). Blacet, Fielding, and Roof suggest a mechanism for the polymerization involving reaction of either CH2CH, HCO, H. or more complex radicals with acro- lein molecules, probably at the C=C bond. The process concerned is probably to be compared with the polymerization of ethylene by methyl radicals (52~. REFERENCES (1) AEEROYD AND NOURISH: J. Chem. Soc. 1936, 890. (2) ALLEN AND SICKMAN: J. Am. Chem. Soc. 66, 1251 (1934~. (3) BAMFORD AND NOURISH: J. Chem. Soc. 1936, 1504. (4) BARAE AND STYLE: Nature 136, 307 (1935~. (5) BLACET, FIELDING, AND ROOF: J. Am. Chem. Soc. 69, 2375 (1937~. (6) BLACET AND ROOF: J. Am. Chem. Soc. 68, 73 (1936~. (7) BLACET AND ROOF: J. Am. Chem. Soc. 68, 278 (1936~. (8) BI.ACET AND VOLMAN: J. Am. Chem. Soc. 60, 1243 (1938~. (9) BLACET, YOUNG, AND ROOF: J. Am. Chem. Soc. 59, 608 (1937~. (10) BOWER AND DE LA PRAUDIERE: J. Chem. Soc. 1934, 1503. (11) BURTON: J. Am. Chem. Soc. 68, 1645, 1655 (1936~; J. Phys. Chem. 41, 322 (1937~. (12) DAMON AND DANIELS: J. Am. Chem. Soc. 66, 2363 (1933~. (13) DIRGE AND KISTIA1~0WSEY: Proc. Nat1. Acad. Sci. U. S. 18, 367 (1932~; Phys. Rev. 46, 4 (1934~. (14) DUNCAN, ELLS, AND NOYES: J. Am. Chem. Soc. 58, 1454 (1936~. (15) GLAZEBROOE AND PEARSON: J. Chem. Soc. 1937, 567. (16) HENS! AND SCROD: Z. Physik 49, 774 (1928~. (17) HERZBERG: Trans. Faraday Soc. 27, 378 (1931~. (18) HERZBERG AND FRANZ: Z. Physik 76, 720 (1932~.
ALDEHYDE AND KETONE PHOTOLYSIS (19) HOWE AND NOYES: J. Am. Chem. soc. 68, 1404 (1936~. (20) KISTIAKOWSBY: CO1d Spring Harbor Symposia QUant. BiO1. 3, 44 (1935~. (21) LEERMAKERS: J. Am. Chem. soc. 56, 1537 (1934~. (22) LEERMAKERS: J. Am. Chem. soc. 66, 1899 (1934~. (23) LEERMAKERS: CO1d Spring Harbor Symposia QUant. Biol. 3, 49 (1935~. (24) LEIGHTON AND BLACET: J. Am. Chem. soc. 64, 3165 (1932~. (25) LEIGHTON AND BLACET: J. Am. Chem. soc. 55, 1766 (1933~. (26) LEIGHTON, LEVANAS, BLACET, AND ROWE: J. Am. Chem. soc. 69, 1843 (1937~. (27) MULLIKEN: J. Chem. PhYS. 3, 564 (1935~. (28) NOURISH: Trans. Faraday soc. 30, 107 (1934~. (29) NOURISH: Acta Physicochim. U.R.S.S. 3, 171 (1935~. (30) NOURISH: Trans. Faraday soc. 33, 1521 (1937~. (31) NOURISH AND APPLEYARD: J. Chem. soc. t933, 874. (32) NOURISH AND BAMFORD: Nature 138, 1016 (1936~. (33) NOURISH, CRONE AND SALTMARSH: J. Chem. soc. 1934, 1456. (34) NOURISH AND KIRRBRIDE: Trans. Faraday sac. 27, 404 (1931~. (35) NOURISH AND Ki~gBR~DE: J. Chem. SOC. 1932, 1518. (36) NONES, DUNCAN, AND MANNING: J. Chem. PhYS. 2, 717 (1934~. (37) PATAT: z. physik. Chem. B26, 208 (1934~. (38) PATAT: z. physik. Chem. Bat, 105 (1935~. (39) PATAT AND LOCKER: Z. physik. Chem. B27, 431 (1935~. (40) PATAT AND SACHSSE: Naturwissenschaften 23, 247 (1935~. (41) PEARSON: Nature 136, 221 (1935~. (42) PEARSON AND GLAZEBROOK: J. Chem. soc. 1936, 1777. (43) PEARSON AND PURCELL: J. Chem. soc. 1936, 1151. (44) PEARSON AND PURCELL: J. Chem. soc. 1936, 253. (45) PRICE: Phys. Rev. 46, 529 (1934~; 47, 444 (1935~; J. Chem. Phys. 3, 256 (1935~. (46) RICE AND HERZFELD: J. Am. Chem. soc. 66, 284 (1934~. (47) RICE, RODOWSKAS, AND LEWIS: J. Am. Chem. soc. 66, 2457 (1934~. (48) ROLLEFSON: 1. Phys. Chem. 41, 259 (1937~. (49) SCHEIBE, POVENZ, AND LINSTROM: z. physik. Chem. B20, 292 (1933~. (50) SPENCE AND WILD: Nature 138, 206 (1936~; J. Chem. soc. 1937, 352. (51) TAYLOR: J. Phys. Chem. 34, 2049 (1930~. (52) TAYLOR AND JUNGERS: Trans. Faraday soc. 33, 1356 (1937~. (53) THOMPSON AND LINNETT: J. Chem. soc. 1936, 1452. Addendum 61 Since the completion of the preceding discussion the results of some un- published work carried out by E. Gorin in Moscow, U.S.S.R., have been communicated privately to the Committee. The newer data may modify some of the preceding conclusions on the problem of mechanism in such photolyses. Gorin has studied the photolysis of acetaldehyde, acetone, formaldehyde and methyl ethyr ketone in the presence of iodine vapor. It has been shown that a few tenths of a millimeter pressure of iodine molecules is sufficient to react with all the free radicals formed in the primary process. 2 Contributed by Hugh S. Taylor.
62 PHILIP A. LEIGHTON From analyses for the products RI and HI so produced, the number of radicals produced by the action of the light can be deduced. With acetaldehydeiodine systems Gorin finds methane and carbon monoxide in addition to methyl iodide. At a given wave length the ratio CH4:CH3I is constant, independent of the temperature and of the pressure of iodine and aldehyde. There are, therefore, on this evidence, two pri- mary processes CH3CHO CH4 + CO (I) CH3 + CHO (II) Gorin concludes that at 3130 A. reaction II is 2.6 times as probable as I, but that at 2537 A. reaction I is 2.9 times more probable. The quantum yield is unity for the formation of CH4 + CH3I at both wave lengths. Gorin finds that HCO radicals do not react with iodine molecules below 100°C., but do react to give H2CO + CO. The formaldehyde formed was equal to half the methyl iodide formed, and equal to the excess of carbon monoxide over methane formed by reaction I. Above 100°C. hydrogen iodide begins to be formed, suggesting that formyl radicals decompose at these temperatures. This gives a minimum value of 26 kg-car. for the activation energy of the reaction HCO = H + CO. The evidence as to the influence of wave length on the alternative modes of decomposition is not in accord with other evidence on the relative amounts obtained from other studies (see page 751~. In the photolysis of acetone in presence of traces of molecular iodine no appreciable carbon monoxide formation was observed by Gorin below 60°C. Methyl iodide is formed, with a quantum yield of unity, at all wave lengths. Acetyl iodide in small amounts and diacetyl were qualita- tively detected. This supports the conclusion that the primary process Is represented entirely by process B of the preceding paper (page 7509. The quantum yield of unity would appear to eliminate theories b and d (page 756) of the discussion by Leighton. With formaldehyde and iodine, at all wave lengths below 3650 A., only carbon monoxide and hydrogen iodide are formed in the ratio 2HI:CO. The primary process should therefore be, exclusively, H2CO + he = H + HCO At 3650 A. hydrogen is also found, indicating the additional process H2CO + he `3650> = H2 + CO (I) (II)
ALDEHYDE AND KETONE PHOTOLYSIS 63 The relative probabilities are 2.5 for reaction I to 1 for reaction II. The ratio is not changed by replacing iodine with mercuric iodide as acceptor. With methyl ethyl ketone and iodine, Gorin's results indicate that not more than a few per cent of the total primary process produces saturated hydrocarbon and carbon monoxide; the radical-producing processes are therefore overwhelmingly predominant. With acetaldehyde and iodine no polymerization was found by Gorin, indicating that it is the free radicals which are largely responsible for the polymerization observed during photo- lysis in the absence of iodine. ..~
